Advances, Systems and Applications
Table 2 Results of mAP evaluation against adversarial attack methods on the MS COCO 2017 test set
From: RPU-PVB: robust object detection based on a unified metric perspective with bilinear interpolation
Method | Conference | Clean | \(A_{cls}\) [21] | \(A_{loc}\) [21] | CWA [19] | DAG [34] |
|---|---|---|---|---|---|---|
SSD [29] | ECCV2016 | 42.0 | 0.4 | 1.8 | 0.1 | 8.1 |
MTD [18] | ICCV2019 | 24.2 -17.8 | 13.0 +12.6 | 13.4+11.6 | 7.7+7.6 | - |
CWAT(PGD-10) [19] | CVPR2021 | 23.7 -18.3 | 14.2 +13.8 | 15.5+13.7 | 9.2+9.1 | - |
RobustDet [20] | ECCV2022 | 36.7 -5.3 | 20.6 +20.2 | 19.4+17.6 | 20.5+20.4 | 24.5 +16.4 |
RobustDet* [20] | ECCV2022 | 36.0-6.0 | 20.0+19.6 | 19.0+17.2 | 19.9+19.8 | 16.6+8.5 |
RPU-PVB | - | 36.2-5.8 | 24.5 +24.1 | 27.1 +25.1 | 25.3 +25.2 | 26.6 +18.5 |